SaaS MRR, churn and retention analysis

Mizan Cloud is a fictional B2B reporting and dashboards product sold to companies in Saudi Arabia and the GCC on monthly and annual subscriptions, with a free 14-day trial and five price plans from Starter to Enterprise. About 600 accounts signed up between January 2023 and December 2025, and monthly recurring revenue (MRR) has grown steadily.

The board is happy with the top line, but the new CFO is not convinced the growth is healthy. Sales report new MRR from their CRM, customer success reports churn from cancellation emails, and the finance team's MRR figure comes from invoices, which mixes annual and monthly billing. Nobody can say how much of the growth comes from new customers versus upgrades, how many customers leave each month, whether recent cohorts retain better than older ones, or why customers cancel.

The subscription table holds the truth, but it needs care. A plan change closes the old row with status superseded and opens a new one on the same day, so not every ended row is churn. Trials are rows with zero MRR. Some customers churn and come back. Five internal test accounts generate heavy usage and must be excluded from every metric. The event stream contains duplicate events from client retries, and the newest cohorts have had little time to churn. Your task is to build the SaaS metrics from first principles so that sales, customer success and finance finally use the same numbers.